Output details

Motion Adaptation With Motor Invariant Theory

Significance: Even though we are capable of complex locomotion shortly after birth, human motion is still poorly understood. Our sophisticated anatomy and lack of clarity around cognitive function in motor control makes defining the processes involved difficult to model. Understanding locomotion has numerous applications, from the simulation of emergent virtual character behaviour in Films and Computer Games, to robotics and medical rehabilitation applications.

Originality: Contemporary advances in robotics, neurobiology and motor control suggests the existence of motion primitives - abstract building blocks which are used to construct complex motion in the same manner as an alphabet forms a sentence. This paper develops a model of human motion based on motion primitive adaptation. Our work points towards a mathematical theory governing the process by which these primitives adapt to changes to the body or the environment. This theory provides the framework for understanding the basic processes governing biological motion by allowing us to describe adaptation mathematically.

Rigour: We validated this theory for motion adaptation by demonstrating that we can enhance the stability of an energy efficient bipedal gait simulation under different environmental conditions. We also demonstrate that our results correspond closely with actual human motion adaptation - an exciting finding which we are eager to develop in later work. These findings provide a better understanding of human motion which has applications in many different scientific and medical disciplines, including realistic virtual character animation, robotics, gait analysis and rehabilitation robotics.